Currently, as requirements on a data transmission rate, communication quality, and the like of mobile communication constantly increase, an existing frequency band used for mobile communication has become extremely congested. However, in a 6-300 GHz millimeter-wave frequency band, a large quantity of spectrum resources have not been allocated for use. Introducing a millimeter-wave frequency band into cellular access communication to make full use of a high bandwidth resource of the millimeter-wave frequency band is one of important research directions of a next-generation 5G (5th Generation) mobile communications technology.
In an existing research, a high frequency band represented by the millimeter-wave frequency band is mainly applied to an indoor short range communications scenario. In an outdoor scenario, because of features such as a complex terrain, a relatively high path loss of the high frequency band, a weak capability of penetrating an obstacle, and a severe rain fade at some frequencies, application of the high frequency band in the outdoor scenario is seriously limited. However, because the high frequency band has a short wavelength and can be easily used to implement a large-scale array antenna, the high frequency band may bring a large directional antenna gain by using a beamforming technology, so as to effectively compensate for the high path loss of the high frequency band. This also provides a possibility for application of the high frequency band in intermediate and long distance transmission in the outdoor scenario.
Generally, a base station applied in a high-frequency communications system is referred to as a high-frequency base station, and a base station applied in a low-frequency communications system is referred to as a low-frequency base station. In the high-frequency communications system, both a high-frequency base station and user equipment can use a large-scale antenna array to perform beamforming. Directional beams (wide beams or narrow beams) of different widths may be formed by adjusting phases or amplitudes of each antenna unit, and/or digital weighted vectors on a plurality of radio frequency (Radio Frequency, RF for short) channels.
Each random access process defined in a current Long Term Evolution (Long Term Evolution, LTE for short) system is based on an omnidirectional beam, but in the high-frequency communications system, a directional beam is needed. Therefore, the LTE random access process cannot be directly used in the high-frequency communications system, and further improvement or a new design needs to be performed according to a feature of a high-frequency directional antenna.
In the existing solution, a random access process in the LTE system is extended to the high-frequency communications system, and transmitting/receiving directional beam pairs are traversed to achieve an omnidirectional antenna transmitting/receiving effect. In addition, an optimal transmission directional beam pair is obtained. After random access fails, a transmit power and a beam width that are used for retransmitting a random access signal may be adjusted to increase a success rate of the random access and extend a coverage area of uplink access.
However, in the high-frequency communications system, the high path loss brought by the high frequency band needs to be compensated for by using a high beam gain brought by an antenna array. The high beam gain is obtained based on beam alignment on a receive end and a transmit end. Once beam misalignment occurs on the receive end and the transmit end, quality of a received signal sharply declines, and normal data communication is interrupted. Therefore, in the high-frequency communications system, to ensure normal data communication, beam training needs to be periodically or aperiodically performed, so that the receive end and the transmit end can use an optimal transmitting/receiving beam pair to perform data transmission. However, in the foregoing prior art, only an initial random access process in the high-frequency communications system is considered, and the initial random access process is not considered in combination with a subsequent beam training process.